Patients suffering from AIDS are at risk of developing significant ocular problems either as a result of the human immunodeficiency virus (HIV) itself or as a result of opportunistic infections. Some of these problems can lead to blindness if left untreated. Among the many pathogens that can lead to blindness, cytomegalovirus (CMV) is by far the most common. In FY 1992 several approaches have been taken to investigate this disease. Some CMV retinitis studies focused on prevention. Early initiation of therapy is important for the preservation of useful vision. However, all the drugs available are virostatic, not virocidal; patients are required to take medication for life if they want to preserve their sight. If one could prevent the development of CMV retinitis, it would be considered a major advance. In this regard, we have participated in a phase I trial of immune globulin in patients at risk for development of CMV retinitis. Other CMV retinitis studies focused on therapy. We have continued to evaluate the ability of foscarnet and ganciclovir to prevent progression of CMV retinitis. Two particular questions evaluated were (1) the possible options in the treatment of patients with end-stage CMV retinitis who have silicone oil in their eyes following retinal detachment and (2) the possibility of using a drug delivery system for the eye, which would obviate the need for systemic therapy. In studies of pediatric AIDS during FY 1992, we continued to evaluate the incidence of ocular infection in about 220 children with AIDS. The incidence of complications is lower in children than in adults. However, CMV retinitis in children is more difficult to treat. In a subpopulation of pediatric patients treated with 2',3'-dideoxyinosine (ddI), peripheral retinal changes suggesting drug toxicity have been observed and reported as our monitoring of these children continues. Other FY 1992 studies on AIDS include evaluation of a new antiviral drug for the treatment of toxoplasmosis in AIDS patients' eyes.